Chloroplast Biogenesis During Rehydration of the Resurrection Plant Xerophyta Humilis: Parallels to the Etioplast–Chloroplast Transition

Plant, Cell and Environment (2008) 31, 1813–1824 doi: 10.1111/j.1365-3040.2008.01887.x

Chloroplast biogenesis during rehydration of the resurrection plant Xerophyta humilis: parallels to the etioplast–chloroplast transition

ROBERT A. INGLE1*, HELEN COLLETT1*, KEREN COOPER1, YUICHIRO TAKAHASHI2, JILL M. FARRANT1 & NICOLA ILLING1

1Department of Molecular and Cell Biology, University of Cape Town, Private Bag, Rondebosch 7701, South Africa and 2The Graduate School of Natural Science and Technology, Okayama University, 3-1-1 Tsushima-naka, Okayama 700-8530, Japan

•- ABSTRACT initially the superoxide radical (O2 ) and singlet oxygen 1 ( O2) (Ivanov & Khorobrykh 2003; Møller, Jensen & De-etiolation of dark-grown seedlings is a commonly used Hansson 2007), and chloroplasts contain several antioxi- experimental system to study the mechanisms of chloro- dant systems to scavenge ROS (Apel & Hirt 2004; Foyer & plast biogenesis, including the stacking of thylakoid mem- Noctor 2005). The equilibrium between ROS production branes into grana, the response of the nuclear-chloroplast and scavenging can be perturbed by environmental stresses transcriptome to light, and the ordered synthesis and leading to a rapid increase in ROS concentration (Apel & assembly of photosystem II (PSII). Here, we present the Hirt 2004; Møller et al. 2007). Under water-deficit stress, xeroplast to chloroplast transition during rehydration of the especially under high-light conditions, the excitation energy Xerophyta humilis resurrection plant as a novel system for harvested by chlorophyll can greatly exceed the demand of studying chloroplast biogenesis, and investigate the role of the Calvin cycle for ATP and NADPH, leading to overre- light in this process. Xeroplasts are characterized by the duction of the electron transport chain and enhanced gen- presence of numerous large and small membrane-bound eration of ROS (Smirnoff 1993; Apel & Hirt 2004; Møller vesicles and the complete absence of thylakoid membranes. et al. 2007). While ROS play critical roles in cell signalling While the initial assembly of stromal thylakoid membranes (Kovtun et al. 2000; Foyer & Noctor 2005), they can also occurs independently of light, the formation of grana is light cause extensive oxidative damage to macromolecules such dependent. Recovery of photosynthetic activity is rapid in as lipids, proteins and nucleic acids (Møller et al. 2007). plants rehydrated in the light and correlates with the light- Resurrection plants, which are able to tolerate the loss of dependent synthesis of the D1 protein, but does not require 95% of protoplasmic water and recover full metabolic de novo chlorophyll biosynthesis. Light-dependent synthe- activity in existing tissues upon rehydration, avoid a toxic sis of the chlorophyll-binding protein Lhcb2 and digalacto- build-up of ROS by a controlled and reversible shutdown of syldiacylglycerol synthase 1 correlated with the formation photosynthesis early on during the drying process (Sherwin of grana and with the increased PSII activity. Our results & Farrant 1998; Farrant 2000). suggest that the molecular mechanisms underlying photo- Angiosperm resurrection plants can be classified into two morphogenic development may also function in desiccation groups based on the mechanisms they utilize to shut down tolerance in poikilochlorophyllous resurrection plants. photosynthesis during desiccation. Homoiochlorophyllous species, such as Craterostigma, retain their chlorophyll and Key-words: desiccation tolerance; photosynthesis; resurrec- rely on pigment production and morphological changes, tion plant; xeroplast. such as leaf folding, to prevent light–chlorophyll interactions during desiccation (Sherwin & Farrant 1998; Farrant 2000). In contrast, poikilochlorophyllous resurrection INTRODUCTION plants, such as Xerophyta, dismantle thylakoid membranes The light reactions of photosynthesis couple the absorption and break down chlorophyll during drying (Tuba et al. 1996; of light by chlorophyll to the generation of energy and Sherwin & Farrant 1998; Farrant 2000). Recent studies have indicated that down-regulation of reducing power to drive the fixation of CO2 (Nelson & Yocum 2006). Operation of the light reactions inevitably photosystem II (PSII) subunit expression also occurs in leads to the formation of reactive oxygen species (ROS), poikilochlorophyllous species during desiccation (Collett et al. 2004; Ingle et al. 2007). PSII is a large protein complex Correspondence: N. Illing. Fax: +27 21 689 7573; e-mail: located predominately in the granal thylakoid membranes [email protected] of the chloroplast, and contains approximately 25 protein *These authors contributed equally to this work. subunits encoded by the psb genes (Mullineaux 2005; © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd 1813 1814 R. A. Ingle et al.

Nelson & Yocum 2006). Six psb genes were previously Determination of RWC identified as desiccation down-regulated in a small-scale Absolute water content (AWC) of leaf samples was calcu- microarray analysis of Xerophyta humilis gene expression. lated using the formula (fresh biomass–dry biomass)/dry These included psbA, which encodes the D1 subunit of the biomass. RWC was calculated using the formula (AWC PSII core complex, and psbO and psbP, which encode com- ¥ 100)/AWC at full turgor (determined after bagging the ponents of the oxygen-evolving complex (OEC). A reduc- control plants overnight after watering). Ten leaf samples tion in protein levels of several PSII subunits in Xerophyta were taken at each time point from each treatment group viscosa at 55% relative water content (RWC) correlated for determination of RWC. with the cessation of photosynthetic activity in this species (Ingle et al. 2007). Upon rehydration, photochemical activity recovers Determination of chlorophyll content rapidly in Xerophyta species (Sherwin & Farrant 1996), The leaf samples were cut into small pieces, and chlorophyll suggesting that they have evolved mechanisms to allow the was extracted in 100% acetone for 4 d at 4 °C. Total rapid biogenesis and assembly of both thylakoid mem- chlorophyll (a + b) content (mg g DW-1) was determined branes and the photosynthetic apparatus. The molecular spectrophotometrically using the equation (7.05 ¥ A661.6) + basis of this process and the signalling events involved (18.09 ¥ A644.8) as described in Lichtenthaler (1987). are unclear, although the role of water availability is ob- vious. Strikingly, similar processes occur in the etioplast– chloroplast transition during photomorphogenesis when Measurement of PSII operating efficiency light acts as the signal for chlorophyll biosynthesis, forma- The quantum yield of photosystem II (FPSII), the proportion and stacking of thylakoid membranes, and translation tion of light absorbed by the PSII antennae used in photo- of several PSII mRNAs including psbA, psbB and psbC chemistry (Genty, Briantais & Baker 1989), was determined (Klein & Mullet 1987; von Wettstein, Gough & Kannangara by measurement of chlorophyll fluorescence using a PAM- 1995; Baena-Gonzalez & Aro 2002). Here, we present the 2100 portable chlorophyll fluorometer (Heinz-Walz GmbH, reassembly of chloroplasts during rehydration of X. humilis Effeltrich, Germany). The leaf samples were light adapted as a novel system to study chloroplast biogenesis, and dem- at a photosynthetic flux of ~50 mmol m-2 s-1 for 15 min prior onstrate the role of light in several key events in this to measurement of FPSII. process.

CO2 measurements MATERIALS AND METHODS The rate of net CO2 assimilation or release was determined Plant material and culture using an LI-6400 portable photosynthesis system (Li-Cor Biosciences, Lincoln, NE, USA), operated at an ambient Xerophyta humilis plants were collected from Borakalalo CO2 concentration of 350 ppm. The parameters A and National Park (Limpopo Province, South Africa), potted Rd were calculated using the equations described by von and grown under glasshouse conditions as described in Caemmerer & Farquhar (1981). Sherwin & Farrant (1996). Prior to this study, the plants were transferred to a controlled environment room with a photosynthetic flux of ~200 mmol m-2 s-1 under a 16 h Chloroplast ultrastructural studies light/8 h dark cycle at 25 °C. The plants were dried down Chloroplast ultrastucture was examined using transmission by withholding water for 2 weeks, and then kept in a electron microscopy as previously described in Cooper & desiccated state for a further 2 weeks prior to rehydra- Farrant (2002). Briefly, small pieces of leaf tissue (approxi- tion. Hydrated (control) plants were regularly watered mately 2 mm2) were excised from the middle of four differ- throughout. Light-excluding boxes were placed over the ent leaves, and RWC was determined for each leaf. Fixation plants the previous evening for the rehydration in the was carried out in 2.5% glutaraldehyde in 0.1 m phosphate dark experiments. buffer (pH 7.4) containing 0.5% caffeine, and samples were postfixed in 1% osmium in phosphate buffer.After dehydra- Rehydration time course tion in a graded ethanol series, the tissue was infiltrated with epoxy resin over 4 d. The samples were embedded in epoxy Desiccated plants were rehydrated under a normal 16 h resin, hardened at 60 °C for 16 h, and sectioned at a gold light/8 h dark cycle or in continuous darkness, beginning 1 h interference colour (95 nm) using a microtome. Sections prior to ‘dawn’. Each rehydration experiment spanned a were stained with 2% uranyl aceate and 1% lead citrate, and 36 h time course as previous studies had suggested a sub- were viewed with a transmission electron microscope. The stantial recovery of PSII activity in X. humilis within this dimensions of chloroplasts, vesicles, plastoglobuli, starch time frame (Sherwin & Farrant 1996). Tissue samples were bodies, thylakoid membranes and grana were measured collected immediately prior to rewatering and at 3, 6, 9, 12, with Image-Pro 6.2 (Mediacybernetics, Bethesda, MD, 15, 18, 24 and 36 h post-watering and from the hydrated USA).Measurements were made on three images per RWC, (control) plants at the same time points. from at least two different leaves, for both light regimes. © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 1813–1824 Chloroplast biogenesis during rehydration of X. humilis 1815

Isolation of partial psbD, psbS and dgd1 gels and was transferred to nitrocellulose membrane. Mem- cDNAs from X. humilis branes were blocked for 2 h in 1 ¥ TBST containing 10% w/v non-fat milk powder. Primary antibodies were diluted in First-strand cDNA was synthesized from 2 mg of total RNA 1 ¥ TBST [digalactosyldiacylglycerol synthase 1 (DGD1) using a SuperScript™ II Reverse Transcriptase system 1:1000, Lhc2b 1:2000 and D1 1:4000] containing 10% w/v (Invitrogen, Carlsbad, CA, USA). One microlitre of the non-fat milk powder. Blots were incubated with primary resulting cDNA was used as the template in a standard antibody for 2 h, followed by 3 ¥ 5 min washes in 1 ¥ TBST 30 mL PCR reaction. PCR primers were designed to con- and incubation with secondary antibody (rabbit IgG HRP, served regions with low amino-acid codon degeneracy 1:5000 dilution) for 1.5 h. Bands were detected using chemi- based on known homologs in the National Center for Bio- luminescence as described by Durrant & Fowler (1994). technology Information (NCBI) databases. The X. humilis psbD (~1 kb) and psbS (~0.5 kb) partial cDNAs were isolated using the following primers psbD:5′-GACTGG RESULTS TTACGRAGGGACCG-3′ and 5′-GGTAGAACCTCCTC CTCATCAGGGA-3′ (annealing temperature, 58 °C), and Chloroplast biogenesis during rehydration of psbS:5′-GTNGGYCGYGTTGCYATG-3′ and 5′-ATNG X. humilis CRGCRANGAAGAAGAA-3′ (annealing temperature, Chloroplast ultrastructure undergoes major modifications 62 °C). A 103 bp sequence from the 3′ end of dgd1 was during rehydration in Xerophyta species. We defined six previously isolated in a differential display PCR screen stages of chloroplast biogenesis during rehydration on the (Collett, unpublished data). An additional 1.4 kb of basis of detailed measurements made of chloroplasts, 5′sequence was isolated by RT-PCR using the degenerate membrane-bound vesicles, plastoglobules, starch granules, primer 5′-ACAACAGCNAGTCTTCCNTGGATG-3′ in assembling thylakoid membranes, single thylakoids and combination with the 3′ gene-specific primer 5′-GAAA grana (summarized in Table 1 and Fig. 1). The membrane- TTGACATTTGTACCTGGC-3′. The resulting PCR bound vesicles could be divided into two populations. The fragments were cloned into the pGEM-T-Easy vector larger vesicles (diameter range, 60–200 nm) were consis- (Promega, Madison,WI, USA) and were sequenced. Nucle- tently less stained by osmium than the smaller vesicles otide sequences were deposited in GenBank (see further (diameter, 20–60 nm). These membrane-bound vesicles dif- discussion for accession numbers). fered from the oval osmophilic plastoglobules (diameter, 25–85 nm), which were present at all stages of chloroplast Northern blot analysis biogenesis (Fig. 1). We have coined the term ‘xeroplast’ to describe the stage 1 chloroplasts that were present in dry Total RNA was isolated from the leaf tissue using leaves. Xeroplasts are characterized by the presence of TriReagent (Molecular Research Centre, Inc., Cincinnati, numerous large and small membrane-bound vesicles, with a OH, USA). For northern blot analysis, 20 mg of total RNA length-to-width ratio (L/W) between 1.0 and 2.0, and a was transferred onto nitrocellulose membrane following chloroplast L/W ratio between 1.0 and 1.6. The smaller formaldehyde gel electrophoresis. Blots were prehybridized membrane-bound vesicles were often clustered together in in buffer containing 50 mm sodium phosphate buffer (pH string-like arrays in xeroplasts (Fig. 1). 6.8), 5 ¥ SSC, 5 ¥ Denhardt’s solution, 50% formamide, Stage 2 of chloroplast biogenesis was characterized by a 0.1% (w/v) sodium dodecyl sulphate (SDS), 0.1% (w/v) change in the size distribution of the membrane-bound sodium pyrophosphate and 50 mg mL-1 salmon sperm vesicles. The more osmophobic larger membrane-bound DNA, and were then probed overnight at 42 °C with 32P- vesicles maintained an L/W ratio between 1.0 and 2.0, while labelled cDNA probes prepared using the Megaprime kit the smaller membrane-bound vesicles started to elongate to (Amersham Pharmacia Biotech, Piscataway, NJ, USA). The an L/W ratio between 2.0 and 7.0.These smaller membrane- blots were washed for 2 ¥ 10 min in 1 ¥ SSC, 0.1% (w/v) bound vesicles were frequently found in a head-to-toe SDS at RT, followed by 2 ¥ 10 min in 0.5 ¥ SSC, 0.1% (w/v) arrangement (Fig. 1). Chloroplasts started to lengthen at SDS at 55 °C, and were exposed to autoradiography film. stage 3 (L/W ratio of >2), and were characterized by the first The following X. humilis cDNA clones were used in north- appearance of a single starch body, the presence of scat- ern blot analysis: psbA (AF545583), psbD (DQ067928), tered larger membrane-bound vesicles (L/W ratio between psbO (DV767869), psbP (AF545584), psbR (AY146990), 1.0 and 2.0) and smaller membrane-bound vesicles (L/W psbS (DQ067929), psbT (DV850415), psbY (DV768147) ratio between 2.0 and 7.0), which were frequently linked and DGD1 (AY186241). together (Fig. 1). The larger membrane-bound vesicles and the plastoglobules were usually found in association with Western blot analysis these strings, which we suggest are the emerging prothylakoid membranes. The larger membrane-bound vesicles Total protein was isolated from the leaf tissue as previously had disappeared by stage 4, and the small, elongated described (Ingle, Smith & Sweetlove 2005). Thirty micro- membrane-bound vesicles were assembled into single thy- grams of total protein was separated on 12% sodium dodecyl lakoid membranes. Starch bodies were now more numerous sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and larger (Fig. 1). Grana were first visible in stage 5 and © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 1813–1824 1816 R. A. Ingle et al.

were assembled from three to four thylakoid membranes (Fig. 1). Further stacking of thylakoids occurred and by stage 6, and grana were more frequent and far thicker, with more than seven thylakoids stacked together (Fig. 1). The populations of chloroplasts for each RWC point during rehydration were categorized according to this staging system to see whether there was a difference in Starch grains chloroplast biogenesis in leaves rehydrated in the dark compared with the light. The rate of change in RWC values during rehydration was not affected by the light regime (Supporting Information Fig. S1). Chloroplast biogenesis following rehydration in X. humilis was rapid, and there was no signiﬁcant difference in the rate of biogenesis from stage 1 through stage 4 for plants rehydrated in the dark com- Dark plastoglobules (diameter range from 25 to 85 nm) 1.2–1.4 Present 1.2–1.4 Present pared with the light (Table 2). The majority of chloroplasts 4

4 had changed from stage 1 to stage 2 within 6 h of rehydra- < > tion, even though the RWC on average was still less than 40%. The leaves of the plants were fully hydrated by 12 h, and contained a mixed population of chloroplasts at different stages of biogenesis. After 12 h, the remaining chloroplasts in the leaves rehydrated in the dark progressed to thylakoids) thylakoids)

Absent 1.2–1.4stage Present 4, but no further (Supporting Information Fig. S2). In Xerophyta humilis contrast, under the 16 h L/8 h D cycle, a few chloroplasts with clear grana (stage 5) were ﬁrst detectable at 12 h following rehydration.The proportions of chloroplasts in stage 5 and stage 6 steadily increased with time following rehydration in the light. 10 nm) Grana <

small vesicles Photosynthetic activity resumes rapidly in the Thylakoids (width presence of light

We used the quantum yield of PSII (FPSII) as a measure of PSII photochemical activity (Genty et al. 1989; Baker &

Rosenquist 2004), and net CO2 assimilation to monitor the recovery of photosynthesis during the rehydration process.

Membrane-bound vesicle (width, 10–40 nm) The recovery of FPSII was rapid under the 16 h L/8 h D cycle, beginning within 12 h of rewatering, and reaching 90% of the control (fully hydrated) plant values at 15 h

(Fig. 2a). In parallel with the recovery in FPSII, net CO2 assimilation was ﬁrst detected at 12 h post-watering

(Fig. 2b). As expected, no CO2 assimilation was observed in the plants rehydrated in constant darkness; however, a Membrane-bound vesicles (width, 20–60 nm) Many Few Few or none modest recovery in FPSII (~25%) was observed in these plants (Fig. 2a).

De novo chlorophyll biosynthesis is not required for resumption of photosynthesis Xerophyta species break down the majority of their chloro-

Unstained membrane- bound vesicles (width, 60–200 nm) phyll during desiccation and resynthesize it upon rewatering (Sherwin & Farrant 1996; Farrant et al. 2003; Supporting Information Fig. S3a). In etioplasts, synthesis of chlorophyl- lide from the precursor protochlorophyllide occurs only in Ultrastructural features that characterize the stages 1 to 6 of chloroplast biogenesis during rehydration in 2.0 1.0–2.0 1.0–2.0 2.0–7.0 Absent Absent 1.2–1.4 Present, few and small 2.0 Few or none Few or none 2.0–7.0 Present, but as strings of 2.0 Absent Absent 2.0–7.0 Present Present (layer of 2.0 Absentresponse Absent to light (von 2.0–7.0 Wettstein et Present al. 1995). Similarly, Mature (layer of no Chloroplast L/W L/W> L/W L/W L/W > > > increase in chlorophyll content was detected in X. humilis plants rehydrated in the dark (Fig. 2c, Supporting Informa- Table 1. Stage L/W, length-to-width ratios. 12 1.0–1.63 1.0–1.6 1.0–2.0 1.0–2.0 1.0–2.0 1.0–2.0 Absent 2.0–7.0 Absent Absent Absent Absent 1.2–1.4 1.2–1.4 Absent Absent 4 5 6 tion Fig. S3b). In contrast, under the 16 h L/8 h D cycle, © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 1813–1824 Chloroplast biogenesis during rehydration of X. humilis 1817

(a) (b)

(e) (f)

(g) (h)

Figure 1. Transmission electron micrographs illustrating ultrastructural features characterizing different stages of chloroplast biogenesis during rehydration in Xerophyta humilis. Stages 1 to 4 of chloroplasts rehydrated in the dark (a,c,e,g) are compared with stages 1 to 4 of chloroplasts rehydrated in the light (b,d,f,h). Stages 5 and 6 were only observed in leaves rehydrated in the light (i,j). Insets are magniﬁed to illustrate assembly of thylakoid membranes. Scale bars = 1 mm. l, large vesicle; o, oval vesicle; e, small elongated vesicle; p, plastoglobule; s, starch grain; t, thylakoid; (i) (j) g, grana.

Table 2. Percentage of chloroplasts in 0 h 3 h 6 h 9 h 12 h 26 h 30 h 34 h 51 h different stages of biogenesis following rehydration in the dark or under a 16 h Dark light/8 h dark cycle Stage 1 100 67 Stage 2 33 75 100 25 Stage 3 25 67 25 Stage 4 8 75 100 100 100 Stage 5 Stage 6 16 h light/8 h dark Stage 1 100 92 25 Stage 2 8 100 75 33 8 Stage 3 17 Stage 4 33 58 58 Stage 5 17 33 33 67 8 Stage 6 33 92

Chloroplasts were staged in 12 independent images, from two different leaves, per time point. See Table 1 for classiﬁcation of stage. chlorophyll content increased to approximately 40% of plants rehydrated in the dark (Fig. 3) during the course of that in the control plants by 36 h post-watering (Fig. 2c). the experiment. However, both psbA and psbD mRNA

However, net CO2 assimilation (and an almost total recov- were detectable in desiccated tissue, suggesting that they ery of FPSII) was detected just after 12 h and prior to any can be stably stored (Fig. 3). increase in chlorophyll content (Fig. 1a–c). This suggests that the residual chlorophyll present in desiccated tissue (<1 mg g DW-1) is sufﬁcient for the resumption of photo- Light is required for translation of the D1 synthetic activity during rehydration. reaction centre protein While psbA mRNA is present in the etioplasts of dark- Water availability is the primary signal for psb grown plants, synthesis of the D1 protein is light dependent gene expression during rehydration (Müller & Eichacker 1999; Zhang & Aro 2002). Given the apparent parallels between the etioplast–chloroplast tran- Down-regulation of psb gene expression occurs at both the sition and chloroplast biogenesis during rehydration, we mRNA and protein level during desiccation in Xerophyta examined D1 protein levels by Western blotting. While species (Collett et al. 2004; Ingle et al. 2007). We thus analy- psbA mRNA was present at all time points including in sed psb transcript levels during rehydration to determine desiccated tissue (Fig. 3), the D1 protein was only detect- whether light is required for transcriptional activation of able from 12 h post-watering in plants rehydrated in the these genes. In addition to the six psb genes we previously presence of light (Fig. 4) and correlated with the resump- identiﬁed as desiccation down-regulated (Collett et al. tion of photosynthetic activity. The modest recovery in 2004), we isolated partial cDNA clones for psbD and psbS FPSII (~25%) observed for plants rehydrated under con- from X. humilis. psbD encodes the D2 subunit required for stant darkness (Fig. 2a) is presumably because of basal the assembly of the PSII core complex (Baena-Gonzalez & levels of the D1 protein, which are not detectable by Aro 2002), while PsbS is involved in non-photochemical Western blot analysis. quenching (Li et al. 2000) and thus might be important during rehydration to limit damage from excess excitation energy.The mRNA abundance of these eight psb genes was Light is required for the synthesis of two other followed during the rehydration process by northern blot chloroplast proteins implicated in PSII activity analysis (Fig. 3). Under a 16 h L/8 h D cycle, mRNA levels of seven of these eight genes had increased by 6 h post- We also investigated whether the synthesis of two proteins, watering and peaked after 9 h, that is, prior to net CO2 Lhcb2 and digalactosyldiacylglycerol synthase 1 (DGD1), assimilation, while psbA mRNA levels increased only at believed to play important roles in the stability and activity 12 h post-watering and peaked at 24 h. A similar pattern of of PSII and in the stacking of thylakoid membranes was gene expression was observed for psbO, psbP, psbR, psbTn light dependent. In addition to their role in light harvesting, and psbY in the plants rehydrated under constant darkness. the LHCII proteins are thought to play a role in the forma- Transcript levels increased to similar maximum levels albeit tion of granal stacks through protein–protein interactions more slowly, peaking at 24 h after rewatering. In contrast, between LHCII and PSII complexes in adjacent thylakoid the increase in psbS expression was reduced, and no membranes (Mullineaux 2005). Western analysis of one of increase in psbA or psbD mRNA levels was detected in the the LHCII proteins (Lhcb2) revealed that its synthesis was © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 1813–1824 Chloroplast biogenesis during rehydration of X. humilis 1819

(a) 16 h L/8 h D Dark 0.9 0.8 0 6 9 12 24 36 6 9 12 24 36 h 0.7 0.6 psbA 0.5 PSII

F 0.4 psbD 0.3 0.2 psbO 010.1 0 psbP 0 6 12 18 24 30 36 h post-watering psbR (b) 6 psbSp

4 –1 psbTn s

m 2

2 psbY 0 CO

mol –2 18S rRNA m –4 Figure 3. Northern analysis of psb gene expression in –6 Xerophyta humilis plants rehydrated under a 16 h light/8 h dark 0 6 12 18 24 30 36 cycle or in constant darkness. Twenty micrograms of total RNA was probed with 32P-labelled partial cDNA probes for eight psb h post-watering genes from X. humilis. 18S rRNA signal intensity indicates equal (c) loading of the RNA samples. 9

8 –1 7

DW respectively (Dörmann & Benning 2003). In addition to

g 6 its role in contributing to the formation of the proton- 5 impermeable bilayer, a small fraction of the DGDG pool 4 is thought to play a critical role in stabilizing PSII and in 3 stabilizing the formation of the LHCII trimers in the light- chlorophyll 2 harvesting antenna (Dörmann et al. 1995; Steffen et al. mg 1 2005). DGD1 is the major enzyme catalyzing the conver- 0 sion of MGDG into DGDG, and a 103 bp fragment of a 0 6 12 18 24 30 36 DGD1 homolog was previously identiﬁed as dehydration up-regulated in X. humilis in a differential display PCR h post-watering screen (Collett, unpublished data). In the present study, Figure 2. Recovery of photosystem II (PSII) quantum yield (a), a 1.4 kb partial cDNA of this gene was obtained using

CO2 assimilation (b) and chlorophyll content (c) in Xerophyta RT-PCR, and the predicted amino acid sequence shows 80 humilis plants rehydrated under a 16 h light/8 h dark cycle (ᮀ) or in constant darkness ("). Relative water content, PSII quantum yield and chlorophyll content are also shown for control plants 16 h L/8hD Dark not subjected to dehydration. The horizontal bar indicates the 16 h light (white)/8 h dark (black) cycle operating in the growth chamber. ᭜, Represents the control hydrated X. humilis grown 0 6 9 12 24 36 6 9 12 24 36 h under the same light/dark cycle conditions. Values indicated are D1 (PsbA) means Ϯ SD (n = 10), and the results shown are for one experiment representative of three. Ponceau S dependent on the presence of light (Fig. 5) and correlated with the formation of granal stacks in plants rehydrated Figure 4. Western analysis of D1 protein levels during rehydration of Xerophyta humilis. Thirty micrograms of total under the 16 h L/8 h D cycle. protein was probed with a D1 antibody. Equal loading of the gel Two galactolipids, monogalactosyldiacylglycerol was veriﬁed by Ponceau S staining of the membrane after protein (MGDG) and digalactosyldiacylglycerol (DGDG), consti- transfer. The results shown are from one experiment tute up to 50 and 20% of lipids in the thylakoid membranes, representative of three. © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 1813–1824 1820 R. A. Ingle et al.

to study chloroplast biogenesis. This species carries out a 16 h L/8 h D Dark controlled and reversible shutdown of photosynthesis early on during dehydration. However, upon rehydration, rapid 0 12 24 36 12 24 36 h reassembly of the thylakoid membranes and photosynthetic apparatus occurs. As is the case in both proplastid development and etioplast maturation, light plays an important Lhcb2 signalling role in this process. The plastids in desiccated X. humilis leaves, which we Figure 5. Western analysis of Lhcb2 protein levels during have termed ‘xeroplasts’, differ substantially from both rehydration of Xerophyta humilis. Thirty micrograms of total proplastids and etioplasts. Proplastids are small spherical protein was probed with a polyclonal antibody to Lhcb2. Equal organelles (0.2–1 mm in diameter), which originate mater- loading of the gel was verified by Ponceau S staining of the nally, and are maintained in an undifferentiated state in the membrane after protein transfer. The results shown are from one experiment representative of three. developing embryo (Mullet 1998; Vothknecht & Westhoff 2001). Etioplasts differentiate from proplastids in seedlings grown in the dark, and are characterized by the presence of well-developed paracrystalline prolamellar bodies with and 78% identity to the rice and Arabidopsis homologs, single thylakoids extending into the stroma (Robertson & respectively (data not shown). Northern analysis confirmed Laetsch 1974). Xeroplasts in X. humilis are larger (0.7–2 mm that DGD1 is up-regulated in X. humilis during desiccation, in diameter) than proplastids, but lack the paracrystalline with transcript levels detectable only at RWC below 30% prolamellar bodies of etioplasts, which form a reserve of (Fig. 6a). DGD1 mRNA was stably stored in desiccated membrane material that is rapidly rearranged into thyla- tissue, and transcript levels decreased within the first 36 h of koid membranes upon exposure to light (Robertson & rewatering irrespective of the presence of light (Fig. 6b). Laetsch 1974). However, Western blot analysis revealed a marked differ- The ultrastructural changes that accompany the develop- ence in DGD1 protein levels between the two treatment ment of proplastids into chloroplasts include changes in groups. Despite the presence of the DGD1 transcript, no chloroplast shape, starch formation, lamellar extension and DGD1 protein was detected in plants rehydrated in the dark. granal development, and have been used to define clear In contrast,DGD1 protein was detected at 6 h post-watering stages of development (Whatley 1974). Stages of the basic in plants rehydrated under the 16 h light/8 h dark cycle, with pathway include (1) a proplastid stage; (2) an amyloplast levels peaking at 9 h. Levels of this protein declined rapidly, stage in which starch granules appear;(3) an amoeboid stage and by 12 h post-watering, no DGD1 protein was detectable in which plastids become elongated and folded;(4) a stage of by immunoblotting. plastid elongation where there is development of perforated stroma lamellae and later incipient grana; and (5) a maturation stage when the aligned lamellae become continuous, DISCUSSION and grana increase in number and depth of stacking Two experimental systems have been previously used to (Whatley 1977). A final stage, gerontoplast, can be defined describe chloroplast development in higher plants, namely which occurs when leaves senesce (Vothknecht & Westhoff the development of proplastids during seed germination 2001).These senescent chloroplasts are characterized by the and the maturation of etioplasts in plants initially grown in presence of large, unstained membrane-bound vesicles, and the dark. Here, we present the xeroplast to chloroplast tran- osmophilic globules occupy much of the interior (Whatley sition in X. humilis during rehydration as a novel system 1974). Xeroplasts resemble these senescent chloroplasts in

(a) 100 63 48 27 6 % RWC

dgd1 mRNA

18S rRNA

Figure 6. Digalactosyldiacylglycerol (b) 16 h L/8 h D Dark synthase 1 (DGD1) expression in Xerophyta humilis during dehydration 0 6 9 12 24 36 6 9 12 24 36 h (a) and rehydration (b). For Northern analysis, 20 mg of total RNA was probed dgd1 mRNA with a 32P-labelled partial dgd1 cDNA probe. Western analysis was carried out on 30 mg of total protein with a DGD1 DGD1 protein polyclonal antibody. © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 1813–1824 Chloroplast biogenesis during rehydration of X. humilis 1821 that they too are filled with unstained membrane-bound Lhcb2 was found to correlate with the formation of grana in vesicles and osmophilic globules (Fig. 1). However, unlike plants rehydrated under the l6 h L/8 h D cycle (Figs 1 & 5). gerontoplasts, xeroplasts are not senescent, and are able to Photosynthetic activity resumed rapidly in plants rehy- differentiate back into chloroplasts upon rehydration. drated under the 16 h L/8 h D cycle, with net CO2 assimila- Although similar ultrastructural changes are seen during tion recorded by 12 h post-watering, correlating with a xeroplast development, they are more rapid and do not ~65% recovery in FPSII (Fig. 2). Interestingly, while net occur in the same order as the proplastid-chloroplast devel- CO2 assimilation did not occur in plants rehydrated in the opmental sequence. We identified six distinct stages in the dark, a modest recovery in FPSII was observed. The rela- xeroplast–chloroplast transition (Fig. 1, Table 1). The first tionship between FPSII and the rate of linear electron change in ultrastructure organization of xeroplasts was flow through PSII can be complicated under environmental observed within 3 h of rehydration when many of the smaller stress, and the proportion of active PSII centres cannot be unstained membrane-bound vesicles start to elongate to determined by measurement of FPSII (Maxwell & Johnson form the precursors to thylakoid membranes (stage 2).Stage 2000). Nonetheless, the partial recovery of FPSII suggests 3 is characterized by the elongation of chloroplasts and by that some degree of assembly of the PSII core complex may the first appearance of starch bodies while thylakoid precur- occur even in the absence of light. Further study is required sor membranes are still being formed (Fig. 1).There is a clear to determine whether this is in fact the case. correlation between the sequential disappearance of the Light is required for chlorophyll biosynthesis during unstained membrane-bound vesicles and the appearance of proplastid development, etioplast–chloroplast transition the prothylakoid membranes, suggesting that in contrast to (Baena-Gonzalez & Aro 2002) and also xeroplast– proplastids (Muhlethaler & Frey-Wyssling 1959;Vothknecht chloroplast transition (Fig. 2c). However, we found that & Westhoff 2001), the thylakoid membranes are not derived the resumption of photosynthesis occurred prior to any from the inner chloroplast membrane.No membrane-bound increase in chlorophyll content (Fig. 2). While the chloro- vesicles remain by stage 4,and immature,perforated stromal phyll content of X. humilis leaves declines dramatically thylakoids are clearly visible. Progression from stage 1 to during dehydration, a residual amount (approximately 10% stage 4 occurs at the same rate in X. humilis rehydrated of that present in fully hydrated plants) is present in desic- under a 16 h L/8 h D cycle or in the dark (Fig. 1, Table 2). cated leaf tissue, and is apparently sufficient for the initial However, while thylakoid reassembly in X. humilis is recovery of photosynthetic activity. Free chlorophyll is a independent of light, the formation of grana is light depen- potent generator of ROS (Hutin et al. 2003), and the dent with only single appressed thylakoids observed in residual chlorophyll present in desiccated tissue may be plants rehydrated in the dark, that is, chloroplast develop- bound in a protein–chlorophyll complex to prevent photo- ment arrests at stage 4 (Fig. 1, Supporting Information oxidative damage. The primary chlorophyll-binding pro- Fig. S2). During photomorphogenesis, light acts as a signal teins in plants are the LHC proteins, but these are not for the formation of thylakoids and granal stacks in etio- detectable in desiccated tissue of Xerophyta species (Fig. 5, plasts (Lopez-Juez & Pyke 2005).This process is mediated at Ingle et al. 2007). An alternative candidate might be the least in part by phytochrome,as Arabidopsis mutants lacking early light-inducible proteins (ELIPS). These proteins can the chromophore phytochromibilin display reduced granal bind chlorophyll and are transiently expressed during formation during de-etiolation (Chory et al. 1989). Interest- de-etiolation or under high-light conditions (Grimm, Kruse ingly, chloroplasts from X. humilis plants rehydrated in the & Kloppstech 1989; Adamska et al. 1999). Interestingly, an dark (Fig. 1) resembled those observed in several dark- ELIP-like protein (dsp 22) has been identified as desicca- grown Arabidopsis constitutive photomorphogenic mutants; tion up-regulated in Craterostigma plantagineum and has the development of chloroplasts in dark-grown cop mutants been shown to associate with PSII protein–pigment com- is similarly stalled at stage 4 with no more than two layers of plexes (Alamillo & Bartels 2001). thylakoid structures being observed (Deng, Caspar & Quail We also investigated whether light was required for the 1991; Kwok et al. 1996). Thus, the light requirement for transcription of eight psb genes encoding subunits of PSII, granal formation is conserved between etioplast-chloroplast because the down-regulation of psb gene expression occurs and xeroplast-chloroplast development. at both the mRNA and protein level during dehydration Grana are not present in bacteria or algae, and are there- (Collett et al. 2004; Ingle et al. 2007). With the exception of fore not essential for oxygenic photosynthesis. It has been psbA and psbD, water was the primary signal for mRNA suggested that their evolution allowed the formation of accumulation, although the rate of synthesis was delayed in larger LHCII complexes in higher plants without restricting the absence of light (Fig. 3). In contrast, while both psbA quinone diffusion (Mullineaux 2005). Accordingly, PSII and psbD mRNA are apparently stably stored in desiccated complexes in unstacked thylakoids have been shown to tissue, no increase in mRNA levels was observed in the have smaller LHC than those in granal stacks (Armond & absence of light (Fig. 3). Interestingly, there was little dif- Arntzen 1977). The LHCII proteins themselves have been ference between the transcriptional activation of nuclear suggested to play a role in the formation of grana via inter- (psbO, P, R, S, Tn and Y) or plastid (psbA and D) encoded actions with PSII complexes and other LHCII proteins in psb genes, suggesting that despite the apparently disorga- adjacent thylakoid membranes (Mullineaux 2005; Standfuss nized state of the xeroplast in desiccated tissue (Fig. 1), it et al. 2005). In X. humilis, light-dependent synthesis of remains transcriptionally competent. © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 1813–1824 1822 R. A. Ingle et al.

While psbA mRNA was constitutively present, suggested that resurrection plants have co-opted aspects of synthesis of the D1 protein was light dependent and was seed desiccation tolerance into their vegetative desiccation first detectable 12 h after rehydration under a 16 h L/8 h D tolerance as evidenced by the expression of ‘seed-specific’ cycle (Fig. 4), correlating with the resumption of net CO2 genes in their vegetative tissues (Illing et al. 2005). We assimilation. PSII complex assembly during the etioplast– propose that in X. humilis, the molecular mechanisms chloroplast transition has been well characterized (Baena- involved in photomorphogenesis may also be utilized in the Gonzalez & Aro 2002; Blomqvist, Ryberg & Sundqvist desiccation tolerance programme, supporting the hypoth- 2006). Many of the PSII protein subunits accumulate in the esis that vegetative desiccation tolerance is based primarily dark including D2, cytochrome b559 and components of the on altered patterns of gene regulation rather than on the OEC (Müller & Eichacker 1999; Baena-Gonzalez & Aro presence of novel genes. 2002). However, accumulation of the D1 polypeptide occurs only after illumination and involves both light-dependent ACKNOWLEDGMENTS translation initiation and stabilization of the D1 protein by binding of chlorophyll a (Müller & Eichacker 1999; Zhang We thank Borakalalo National Parks for allowing plant & Aro 2002). In addition, there is evidence to suggest that a collection, and are grateful to the following people for precomplex containing D2 and cytochrome b559 is required the donation of antibodies: D1, Eva Aro (University of to act as an acceptor for the elongating D1 polypeptide Turku, Finland), and DGD1, Peter Dörmann (Max Planck ensuring co-translational incorporation of the protein into Institute for Molecular Plant Physiology, Germany). This PSII (Müller & Eichacker 1999).While it is unclear whether research was funded from grants provided by the Univer- the incorporation of D1 into PSII during rehydration in X. sity of Cape Town and the National Research Foundation of humilis requires the presence of a D2/cytochrome b559 South Africa. acceptor, the light dependence of D1 protein synthesis suggests that the PSII reassembly may occur by the same route REFERENCES as that of the etioplast–chloroplast transition. Light was also found to play a role in the post- Adamska I., Roobol-Bóza M., Lindahl M. & Andersson B. (1999) transcriptional regulation of DGD1, which catalyzes the Isolation of pigment-binding early light-inducible proteins from synthesis of the galactolipid DGDG. DGD1 protein was pea. European Journal of Biochemistry 260, 453–460. Alamillo J.M. & Bartels D. (2001) Effects of desiccation on photo- transiently present at 6 and 9 h post-watering in plants rehy- synthesis pigments and the ELIP-like dsp 22 protein complexes drated under the 16 h L/8 h D cycle, but no DGD1 protein in the resurrection plant Craterostigma plantagineum. Plant was detected in plants rehydrated in the dark (Fig. 6b). Science 160, 1161–1170. While light-responsive DGD1 expression has not been pre- Apel K. & Hirt H. (2004) Reactive oxygen species: metabolism, viously reported, a transient peak in mRNA and protein oxidative stress, and signal transduction. Annual Review of Plant levels of MGDG synthase has been previously reported Biology 55, 373–399. in cucumber during de-etiolation (Yamaryo et al. 2003). Armond P.A. & Arntzen C.J. (1977) Localization and characterization of photosystem II in grana and stroma lamellae. Plant Physi- MGDG is the major galactolipid of the thylakoid mem- ology 59, 398–404. brane and also a substrate for the DGD1 enzyme. Baena-Gonzalez E. & Aro E.M. (2002) Biogenesis, assembly and During thylakoid biosynthesis in proplastids, MGDG turnover of photosystem II units. Philosophical Transactions of and DGDG are synthesized on the inner and outer chloro- the Royal Society, London B Biological Science 357, 1451–1459. plast membranes, and are transported to the developing Baker N.R. & Rosenquist M. (2004) Applications of chlorophyll thylakoids by the vesicle-inducing plastid protein VIPP1 fluorescence can improve crop production strategies: an exami- (Dörmann & Benning 2003). Specific fractions of the nation of future possibilities. Journal of Experimental Botany 55, 1607–1621. DGDG pool are also thought to play a role in the stabilizing Blomqvist L.A., Ryberg M. & Sundqvist C. (2006) Proteomic PSII (Dörmann et al. 1995; Steffen et al. 2005; Sakurai et al. analysis of the etioplast inner membranes of wheat (Triticum 2007) and the trimerization of LHCII proteins that form the aestivum) by two-dimensional electrophoresis and mass spec- light-harvesting antenna of PSII (Nussberger et al. 1993; Liu trometry. Physiologia Plantarum 128, 368–381. et al. 2004; Holzl et al. 2006).As the stockpile of MGDG and von Caemmerer S. & Farquhar G.D. (1981) Some relationships DGDG in the membrane-bound vesicles in xeroplasts is between the biochemistry of photosynthesis and the gas exchange of leaves. Planta 153, 376–387. apparently sufficient for thylakoid assembly (irrespective of Chory J., Peto C.A., Ashbaugh M., Saganich R., Pratt L. & Ausubel the presence of the DGD1 protein), the DGD1 produced in F. (1989) Different roles for phytochrome in etiolated and green the light may serve to synthesize DGDG that plays a role in plants deduced from characterization of Arabidopsis thaliana the stabilization of the PSII–LHCII supercomplexes. mutants. The Plant Cell 1, 867–880. In summary, we present the xeroplast–chloroplast transi- Collett H., Shen A., Gardner M., Farrant J.M., Denby K.J. & Illing tion as a novel system for studying chloroplast biogenesis in N. (2004) Towards transcript profiling of desiccation tolerance in higher plants. While the ultrastructural changes that occur Xerophyta humilis: construction of a normalized 11 k X. humilis cDNA set and microarray expression analysis of 424 cDNAs in in the chloroplast differ from those in proplastids or etio- response to dehydration. Physiologia Plantarum 122, 39–53. plasts, there are also striking similarities, with light being Cooper K. & Farrant J.M. (2002) Recovery of the resurrection required for grana formation, chlorophyll biosynthesis, and plant Craterostigma wilmsii from desiccation: protection versus the synthesis of D1 and Lhcb2. It has been previously repair. Journal of Experimental Botany 53, 1805–1813. © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 1813–1824 Chloroplast biogenesis during rehydration of X. humilis 1823

Deng X.W., Caspar T. & Quail P.H. (1991) cop1: a regulatory locus genes is necessary for repression of photomorphogenesis in the involved in light-controlled development and gene expression in dark. Plant Physiology 110, 731–742. Arabidopsis. Genes & Development 5, 1172–1182. Li X.-P., Bjorkman O., Shih C., Grossman A.R., Rosenquist M., Dörmann P. & Benning C. (2003) Galactolipids rule in seed plants. Jansson S. & Niyogi K.K. (2000) A pigment-binding protein Trends in Plant Science 7, 112–118. essential for regulation of photosynthetic light harvesting. Dörmann P., Hoffmann-Benning S., Balbo I. & Benning C. (1995) Nature 403, 391–395. Isolation and characterization of an Arabidopsis mutant defi- Lichtenthaler H.K. (1987) Chlorophylls and carotenoids, the pig- cient in the thylakoid lipid digalactosyl diacylglycerol. The Plant ments of the photosynthetic biomembranes. Methods in Enzy- Cell 7, 1801–1810. mology 148, 1275–1286. Durrant I. & Fowler S. (1994) Chemiluminescent detection systems Liu Z., Yan H., Wang K., Kuang T.,Zhang J.,Gui L., An X. & Chang for protein blotting. In Protein Blotting (ed. B. Dunbar), pp. W. (2004) Crystal structure of spinach major light-harvesting 141–152. Oxford University Press, Oxford, UK. complex at 2.72 Å resolution. Nature 6980, 287–292. Farrant J.M. (2000) A comparison of mechanisms of desiccation Lopez-Juez E. & Pyke K.A. (2005) Plastids unleashed: their devel- tolerance among three angiosperm resurrection plant species. opment and integration in plant development. International Plant Ecology 151, 29–39. Journal of Developmental Biology 49, 557–577. Farrant J.M., Bartsch S., Loffell D., VanderWilligen C. & Whittaker Maxwell K. & Johnson G.N. (2000) Chlorophyll fluorescence – a A. (2003) An investigation into the effects of light on the desic- practical guide. Journal of Experimental Botany 51, 659–668. cation of three resurrection plants species. Plant, Cell & Envi- Møller I.M., Jensen P.E. & Hansson A. (2007) Oxidative modifica- ronment 26, 1275–1286. tions to cellular components in plants. Annual Review of Plant Foyer C.H. & Noctor G. (2005) Redox homeostasis and antioxidant Biology 58, 459–481. signaling: a metabolic interface between stress perception and Muhlethaler K. & Frey-Wyssling A. (1959) Entwicklung und struk- physiological responses. The Plant Cell 17, 1866–1875. tur der proplastiden. Journal of Biophysical and Biochemical Genty B., Briantais J.M. & Baker N.R. (1989) The relationship Cytology 6, 507–512. between the quantum yield of photosynthetic electron transport Müller B. & Eichacker L.A. (1999) Assembly of the D1 precursor and quenching of chlorophyll fluorescence. Biochimica et Bio- in monomeric photosystem II reaction center precomplexes pre- physica Acta 990, 87–92. cedes chlorophyll a-triggered accumulation of reaction center II Grimm B., Kruse E. & Kloppstech K. (1989) Transiently expressed in barley etioplasts. The Plant Cell 11, 2365–2378. early light-inducible thylakoid proteins share transmembrane Mullet J.E. (1998) Chloroplast development and gene expression. domains with light-harvesting chlorophyll binding proteins. Annual Review of Plant Physiology and Plant Molecular Biology Plant Molecular Biology 13, 583–593. 39, 475–502. Holzl G., Witt S., Kelly A.A., Zahringer U., Warnecke D., Dormann Mullineaux C.W. (2005) Function and evolution of grana. Trends in P. & Heinz E. (2006) Functional differences between galactolip- Plant Science 10, 521–525. ids and glucolipids revealed in photosynthesis of higher plants. Nelson N. & Yocum C.F. (2006) Structure and function of photo- Proceedings of the National Academy of Sciences of the United systems I and II. Annual Review of Plant Biology 57, 521–565. States of America 103, 7512–7517. Nussberger S., Dorr K., Wang D.N. & Kuhlbrandt W. (1993) Lipid– Hutin C., Nussaume L., Moise N., Moya I., Kloppstech K. & protein interactions in crystals of plant light-harvesting complex. Havaux M. (2003) Early light-induced proteins protect Arabi- Journal of Molecular Biology 234, 347–356. dopsis from photooxidative stress. Proceedings of the National Robertson D. & Laetsch W.M. (1974) Structure and function of Academy of Sciences of the United States of America 100, 4921– developing barley plastids. Plant Physiology 54, 148–159. 4926. Sakurai I., Mizusawa N., Wada H. & Sato N. (2007) Digalactosyl- Illing N., Denby K.J., Collett H., Shen A. & Farrant J.M. (2005) The diacylglycerol is required for stabilization of the oxygen- signature of seeds in resurrection plants: a molecular and physi- evolving complex in photosystem II. Plant Physiology 145, ological comparison of desiccation tolerance in seeds and veg- 1361–1370. etative tissues. Integrative and Comparative Biology 45, 771–787. Sherwin H.W. & Farrant J.M. (1996) Differences in rehydration of Ingle R.A., Smith J.A.C. & Sweetlove L.J. (2005) Responses to three desiccation-tolerant angiosperm species. Annals of Botany nickel in the proteome of the hyperaccumulator plant Alyssum 78, 703–710. lesbiacum. Biometals 18, 627–641. Sherwin H.W. & Farrant J.M. (1998) Protection mechanisms against Ingle R.A., Schmidt U.G., Farrant J.M., Thomson J.A. & Mundree excess light in the resurrection plants Craterostigma wilmsii and S.G. (2007) Proteomic analysis of leaf proteins during dehydra- Xerophyta viscosa. Plant Growth Regulation 24, 203–210. tion of the resurrection plant Xerophyta viscosa. Plant, Cell & Smirnoff N. (1993) The role of active oxygen in the response of Environment 30, 435–446. plants to water deficit and desiccation. New Phytologist 125, Ivanov B. & Khorobrykh S. (2003) Participation of photosynthetic 27–58. electron transport in production and scavenging of reactive Standfuss J., Terwisscha van Scheltinga A.C., Lamborghini M. & oxygen species. Antioxidants & Redox Signalling 5, 43–53. Kühlbrandt W. (2005) Mechanisms of photoprotection and non- Klein R.R. & Mullet J.E. (1987) Control of gene expression during photochemical quenching in pea light-harvesting complex at higher plant chloroplast biogenesis. Protein synthesis and tran- 2.5 Å resolution. EMBO Journal 24, 919–928. script levels of psbA, psaA-psaB, and rbcL in dark-grown and Steffen R., Kelly A.A., Huyer J., Dormann P. & Renger G. (2005) illuminated barley seedlings. Journal of Biological Chemistry Investigations on the reaction pattern of photosystem II in 262, 4341–4348. leaves from Arabidopsis thaliana wild type plants and mutants Kovtun Y., Chiu W.-L., Tena G. & Sheen J. (2000) From the cover: with genetically modified lipid content. Biochemistry 44, 3134– functional analysis of oxidative stress-activated mitogen- 3142. activated protein kinase cascade in plants. Proceedings of the Tuba Z., Lichtenthaler H.K., Csintalan Z., Nagy Z. & Szente K.

National Academy of Sciences of the United States of America 97, (1996) Loss of chlorophylls, cessation of photosynthetic CO2 2940–2945. assimilation and respiration in the poikilochlorophyllous plant Kwok S., Piekos B., Misera S. & Deng X. (1996) A complement of Xerophyta scabrida during desiccation. Physiologia Plantarum ten essential and pleiotrophic Arabidopsis COP/DET/FUS 96, 383–388.

Vothknecht U.C. & Westhoff P. (2001) Biogenesis and origin of Figure S1. Recovery of relative water content (RWC) of thylakoid membranes. Biochimica et Biophysica Acta 1541, leaves used in electron micrographs characterizing changes 91–101. in ultrastructure associated with chloroplast biogenesis von Wettstein D., Gough S. & Kannangara C.G. (1995) Chlorophyll during Xerophyta humilis rehydration. X. humilis plants biosynthesis. The Plant Cell 7, 1039–1057. Whatley J. (1974) Chloroplast development in primary leaves of were rehydrated under a 16 h light/8 h dark cycle (ᮀ) or in Phaseolus vulgaris. New Phytologist 73, 1097–1110. constant darkness ("). Whatley J. (1977) Variations in the basic pathway of chloroplast Figure S2. Xeroplasts halted at stage 4 of chloroplast devel- development. New Phytologist 78, 407–420. opment when rehydrated in the dark for different lengths of Yamaryo Y.,Kanai D., Awai K., Shimojima M., Masuda T.,Shimada time. Scale bars = 1 mm. H., Takamiya K.-I. & Ohta H. (2003) Light and cytokinin play a Figure S3. Chlorophyll biosynthesis during rehydration. co-operative role in MGDG synthesis in greening cucumber Xerophyta humilis plants during rehydration under a 16 h cotyledons. Plant & Cell Physiology 44, 844–855. Zhang L. & Aro E.M. (2002) Synthesis, membrane insertion and light/8 h dark cycle (a). Comparison of plants rehydrated assembly of the chloroplast-encoded D1 protein into photosys- under 16 h light/8 h dark or under constant darkness for tem II. FEBS Letters 512, 13–18. 48 h (b).

Received 7 August 2008; received in revised form 22 August 2008; Please note: Wiley-Blackwell are not responsible for the accepted for publication 24 August 2008 content or functionality of any supporting materials supplied by the authors. Any queries (other than missing SUPPORTING INFORMATION material) should be directed to the corresponding author Additional Supporting Information may be found in the for the article. online version of this article: